Research On Topology Optimization And Dimensioning Of 5G-based MEC-VANET Networks

The upcoming 5G era mainly focuses on three major application scenarios:Enhanced Mobile Broadband,Ultra Reliable&Low Latency Communication,and Massive Machine-Type Communications.Vehicular Ad-Hoc Network(VANET)is regarded as 5G typical vertical industry,and it is widely recognized that the deep integration with 5G would render better driving experiences for vehicular users.5G is expected to quench the thirst for ultra-high data rate communication and to strengthen the coordination between the intelligent vehicles and the smart road.However,the traditional vehicular cloud computing cannot meet soaring performance requirements of 5G.To this end,this thesis investigates an MVN Architecture.Instead of the conventional RoadSide Units,MVN places 5G Micro Base Station to offer vehicles reliable and real-time services.Then MEC is introduced to preprocess the delay-sensitive data on the edge of the network,thus shortens the delay and saves the bandwidth resources.This thesis studies the TOD problem under MVN architecture to support the VANET's integration with 5G and MEC.First,a tree-based topology for MVN architecture is investigated to render fast-response and reliable communications for VANET and to save the bandwidth resources.Furthermore,this thesis studies the TOD problem for MVN architecture and could provide guidance for MVN architecture's real-world deployment.Next,this thesis formulates the TOD problem as an Integer Linear Program(ILP)and establishes a mathematical model to minimize the total placement cost under the constraints of the topology,communication range and so on.Gurobi solver is applied to verify the feasibility and scalability of TOD optimization.Finally,regarding to Gurobi's issue of low-efficiency in solving the large-size scenarios,MGA and MHA are proposed to improve the computation efficiency.A series of experimental scenarios are designed and the results demonstrate the superiority of MHA.Based on MHA,MCRAA is proposed to conduct the joint optimization of topology and resource allocation.Diverse kinds of MEC are deployed in the diverse areas to save MEC resources and further reduce the deployment cost.